A Non-data aided Algorithm for the Symbol Rate Estimation of CPM Signals

In this paper, a new non-data aided algorithm of the symbol rate of CPM signals is proposed. It is based on the observation that the discrete lines and the modulation parameters are mainly related on the carrier frequency -profile. This algorithm employs the frequency smoothing method to get the estimation of the cyclic spectrum, and adopts nonlinear filtering to extract discrete lines on the cyclic profile. According to the characteristics of cyclic spectrum, single-h signals and multi-h ones are distinguished, and the symbol rate of CPM signals is accurately estimated. Simulation results show that the algorithm can be implemented without precise estimation of carrier frequency. In addition, it has good noise resistant performance, and can be applied to frequency-selective fading channels.